Chloride Selective Calix[4]arene Optical Sensor Combining Urea

Jun 8, 2006 - Benjamin Schazmann , Nameer Alhashimy , and Dermot Diamond *. Contribution from the Adaptive Sensors Group, National Centre for Sensor R...
0 downloads 11 Views 128KB Size
Published on Web 06/08/2006

Chloride Selective Calix[4]arene Optical Sensor Combining Urea Functionality with Pyrene Excimer Transduction Benjamin Schazmann, Nameer Alhashimy, and Dermot Diamond* Contribution from the AdaptiVe Sensors Group, National Centre for Sensor Research, School of Chemical Sciences, Dublin City UniVersity, Dublin 9, Ireland Received March 21, 2006; E-mail: [email protected]

Abstract: A neutral 2-site chloride selective compound has been developed (3), based on a 1,3-alternate tetrasubstituted calix[4]arene providing a preorganized supramolecular scaffold. The resultant supramolecular cavity is among the first to combine urea functional groups bridged with single methylene spacers to pyrene moieties. It combines a naturally and synthetically proven H-bonding system with the elegant ratiometric fluorescent signaling properties of an intramolecular pyrene excimer system, triggered by conformational changes upon anion coordination. The excimer emission of 3 is quenched, with a simultaneous rise in the monomer emission solely by the chloride anion among a wide variety of anions tested. 3 has an association constant of 2.4 × 104 M-1 with chloride. The suitability and advantages of ratiometric optical sensor compounds like 3 for use in practical sensor devices is discussed. 3 has an LOD of 8 × 10-6 M with chloride in acetonitrile-chloroform (95:5 v/v). A dynamic fluorescence study revealed a response time of 300 nm) UV light. This spot was not present in the TLC of 2. Furthermore, the appearance of an excimer emission at 452 nm (λmax) was also seen only in the crude mixture of 3 by fluorescence screening. This was the first evidence that a compound with intramolecular pyrene interaction was present. Following initial workup of the reaction mix, a crude mixture containing 37% 3 (of total peak areas) was present as shown by HPLC analysis.

The crude workup products from the synthesis of 1, 2, and 3 showed very poor solubility in most common solvents, marginally better solubility in MeOH and ACN and good solubility in DMSO or DMF. The preference for highly polar solvents is probably dictated by the polar urea groups and salts present. This affected the ease of purification of 2 and 3 in particular as chromatographic methods had to be used to achieve purity >95%. For 3, an efficient semipreparative HPLC method proved essential. This instrumental approach, previously developed in our group for supramolecule isolation, is of particular benefit when dealing with complex mixtures and low yielding reactions.77 The method used was a fast efficient means of purifying mg quantities of a product using a scaled up analytical HPLC method based on widely available analytical instrumentation. A single pure peak remained as verified by HPLC, revealing an overall product purity of 98% (Figure S1). From the 1H NMR spectrum of 3, there is a single peak for the methylene groups of the calix[4]arene annulus. This is indicative of a 1,3-alternate structure as depicted in Scheme 1.1 The attachment of 4 sterrically bulky appendages like 2 to a calix[4]arene platform may result in a deviation from the more common cone conformation to a less hindered 1,3-alternate configuration. This deviation from the cone conformations is particularly feasible as the calix[4]arene benzene groups are free to rotate through the central aromatic cavity, given the absence of the commonly present upper rim tertiary butyl groups. The two urea protons of start material 2 appeared deshielded from 6.3 and 6.8 ppm into the aromatic region for 3. This signals a large change in the chemical environment of the urea protons in 3. This deshielding of urea protons may be due to proximal pyrene moieties and increased inter- or intramolecular Hbonding of the urea groups of 3. A 1H NMR temperature degradation study of 1 and 2 revealed that degradation generally occurred at temperatures of 80 °C and above. An increased number of peaks and an increased integration numbers for aromatic protons suggested cleavage of the bulky pyrene moiety (Figure S2). The consequences of (77) Schazmann, B.; McMahon, G.; Nolan, K.; Diamond, D. Supramol. Chem. 2005, 17, 393-399. J. AM. CHEM. SOC.

9

VOL. 128, NO. 26, 2006 8609

Schazmann et al.

ARTICLES

Figure 1. Emission spectra of 2 and 3 (1 × 10-6 M) in acetonitrilechloroform (95:5 v/v) showing monomer maxima at 376 nm and 398 nm. Only 3 shows an additional large band at 452 nm due to excimer formation. (The excitation wavelength was 340 nm.)

these findings were that a maximum temperature of only 70 °C for the subsequent synthesis of 3 could be used. This coupled with the solubility and purification issues discussed above led to a final yield of only 2%. The excitation spectrum of 1, 2, and 3 revealed a λmax of 340 nm as an ideal excitation wavelength. Figure 1 shows the emission spectrum of 2 and 3 in acetonitrile-chloroform (95:5 v/v). 2 reveals two monomer emission peaks at 376 nm and 398 nm (λmax). There are no significant emission features >430 nm, indicating the absence of excimer formation. 3 shows the monomer emissions as well as a much larger broad emission band at 452 nm (λmax), characteristic of an intramolecular pyrene excimer emission. Only the pure fraction of 3, unlike the fluorescence spectra of 1, 2 and all the other fractions collected during the purification of 3 by semipreparative HPLC revealed the characteristically broad pyrene excimer emission. For 3, the ratio of excimer (452 nm) to monomer (398 nm) emission remained unchanged at 4.4 in the concentration range 1 × 10-7 to 1 × 10-5 M 3. This further confirmed the presence of pyrene units interacting by an intramolecular mechanism, not an intermolecular one. The ditopic chromoionophore 3 is built on a calix[4]arene platform, lending preorganization to the overall host. Four urea groups providing eight possible H-bonds for anion binding are in close proximity to pyrene moieties, whose orientation relative to each other is thought to change upon guest inclusion. Such a binding event is thus monitored by ratiometric changes in the emission spectrum (Excimer:Monomer ratio) of 3. The changes in the emission spectrum of 3 were examined when screened with 11 common anions. These spanned a comprehensive range of sizes and shapes. 100 equiv of the tertiary butylammonium salt of each anion was added to 1 × 10-6 M solutions of 3 in acetonitrile-chloroform (95:5 v/v). The change in excimer and monomer emission was monitored and the results tabulated in Table 1. Remarkably, only chloride caused a dramatic change in the emission spectrum of 3. There is a sharp decline in excimer emission with a corresponding increase in monomer emission. These observations suggest that the chloride anion appears to selectively coordinate with the urea protons in the cavity of 3 so as to “unstack” or lever apart, the facing π-π stacked pyrenes. This conformational change is depicted in Scheme 2. 8610 J. AM. CHEM. SOC.

9

VOL. 128, NO. 26, 2006

[The structures of free and chloride complexed 3 were created using MM2 force field energy minimization. The energy was reduced to a minimum RMS gradient of 0.100. The software used was Chem 3D Ultra 8.0 supplied by Cambridge Scientific Computing, Inc.] The “unstacking” of the pyrene moieties is justified by the strong reduction of the excimer emission spectrum upon complexation of chloride. The excimer emission of 452 nm (λmax) signifies a considerable blue shift compared to other intramolecular pyrene excimer systems, which typically show a λmax of 480 nm.57 The comparison of the pyrene excimer emission of sandwich-like systems (full overlap) with partially overlapping systems is normally explained in this way.78,79 Sandwich-like systems are described as dynamic excimers, in that pyrene moieties are free to fully overlap. Partially overlapping pyrenes are described as static because some force is inducing a partial overlap. In the case of free 3, there is a strong likelihood of H-bonding acting between urea groups in addition to steric factors, thereby offering a plausible explanation for the observed partially overlapped static excimer. To further support the mode of binding proposed, selected tetrabutylammonium anions were added in excess to 3 in deuterated acetone (Figure S3). Only the chloride salt caused a downfield shift of urea protons, indicative of an H-bonding interaction. Upon addition of 300 equiv. of Cl-, the two triplet signals for the urea protons appear shifted downfield, clear of the main aromatic region at 8.69 ppm and 8.71 ppm. This shows clearly that chloride ions form an inclusion complex within the cavities of 3 involving urea protons. No such discernible change was observed for the smaller fluoride and the larger bromide anions, confirming chloride selectivity. These results mirror the findings of the fluorescence study, where a response appeared exclusive to Cl-. Furthermore, the aromatic protons of 3 appear considerably deconvoluted by the addition of Cl-, suggesting a more symmetrical complex structure compared to the free Host. The separation of previously π-stacked pyrene moieties appears to reduce the influence of the π-electron clouds from the planes of the pyrene moieties on 1H NMR peak splitting and chemical shifts. It appears that a specific cavity effect controls the anion binding characteristics of 3. A ‘lock and key’ or ‘best fit’ model appears to yield an energetically favorable host-guest interaction for chloride in this case. This is in contrast to a scenario where an anion host offers little preorganization, where selectivity is typically dictated by anion basicity.15 When 0-500 equiv of chloride are added to 3, the change in emission can be followed and is shown in Figure 2. At 412 nm there is an isoemissive point which indicates that only one type of complexing mechanism (equivalent in both cavities of 3) is at play perturbing the excimer emission of 3.72 From the change in excimer-to-monomer ratios observed with chloride added, an association constant of 2.4 × 104 M-1 was obtained [The association constant was calculated by non-linear regression analysis and the fitting of experimental data with a standard fluorescence equation by minimizing the sum of squared residuals using the Microsoft Excel add-in SOLVER.

(78) Winnik, F. M. Chem. ReV. 1993, 93, 587-614. (79) Yang, J. S.; Lin, C. S.; Hwang, C. Y. Organic Lett. 2001, 3, 889-892.

Chloride Selective Calix[4]arene Optical Sensor

ARTICLES

Table 1. Fluorescence Changes (I-I0) for 3 upon Addition of 100 equiv. of Specified Aniona fluorescence change (I−I0) λem (nm)

Cl-

F-

Br-

I-

NO3-

Cl04-

AcO-

SCN-

CN-

HSO4-

H2PO4-

H2 O b

398 452

+270 -591

-7 +2

0 +5

-6 -1

-6 -4

-1 +6

+2 -4

-8 0

+2 -8

0 -5

+6 -3

-36 +36

a Conditions: 3, 1.6 × 10-6 M in acetonitrile-chloroform (95:5 v/v), excitation at 340 nm. I : fluorescence emission intensity of free 3. I: fluorescence o emission intensity of 3 with 100 equiv of specified anion in the form of tertiary butylammonium salts. b 1000 equiv added.

Scheme 2 Binding of Chloride Ions by 3. Quenching of Excimer Emission (452 nm) Caused by a Perturbation of the Pyrene π-π Interaction by the Conformational “unstacking” of Pyrene Moieties

The association constants of well-known crown and cryptand alkali metal hosts are typically about 106 M-1 and 1010 M-1 respectively.10 It appears that selective anion hosts seldom reach association constants of this magnitude and are in fact lower by several orders of magnitude, due to factors competing with anion complexation as discussed in the Introduction. Additional factors competing with anion coordination specific to 3 may include the cost in energy to separate the overlapping pyrene π-π bonding systems. The sterically bulky nature of pyrene moieties also aids in the discrimination between anions. In addition, as each equivalent di-urea cavity of 3 binds a chloride anion, the resultant repulsion between these two ions of same charge within 3 may further lower the overall association constant. However, in general, the net effect of the competing factors responsible for low association constants may also be contributing factors for good selectivity. The selectivity of 3 toward chloride can also be explained classically by a ‘best fit’ or “lock and key” model. It is most likely several cumulative factors that contribute to what is ultimately the most important parameter of a sensor design: Selectivity. Few works discussing the combination of urea anion binding sites with pyrenes were found in the literature. Sasaki synthesized a tripodal anion host with pyrenes directly adjacent to 3 thiourea groups.75 In terms of selectivity, Sasaki did not observe a unique selectivity but some deviation from a typical selectivity order, following anion basicity. Werner describes nucleotide anion hosts which have relatively long propyl spacers between pyrene moieties and ureas.76 The separation of π-stacked pyrenes and simultaneous anion binding could not be observed.

Figure 2. Changes of the fluorescence spectrum of a 1 × 10-6 M molar solution of 3 in acetonitrile-chloroform (95:5 v/v) upon addition of the specified number of equivalents of chloride ([Cl-]/[3]). R0: Ratio of excimer (λem ) 452 nm) to monomer (λem ) 398 nm) of free 3. R: Ratio of excimer to monomer with varying [Cl-]. (The excitation wavelength was 340 nm.)

A developer of practical fluorescent probes or sensors can choose from three broad fluorescence signal types.80 Intensity based probes rely on the change of intensity of single wavelengths. The biggest disadvantage here is that for accurate sensing, the exact probe host concentration must be known. Typically, factors like host degradation or leaching into the sample result in ever changing host concentration. Other factors like sample turbidity, intensity of incident light, scattering, innerfilter effects and photo bleaching strongly affect the signal reliability. These disadvantages are largely absent for the other two categories, lifetime based and wavelength-ratiometric fluorescent sensing methods. 3 belongs to the latter class. It is apparent from the literature that there are quite a number of fully characterized fluorescent sensor compounds available (80) Lakowicz, J. R. Topics in Fluorescence Spectroscopy; Plenum Press: New York, 1994; Vol. 4. J. AM. CHEM. SOC.

9

VOL. 128, NO. 26, 2006 8611

Schazmann et al.

ARTICLES

with elegant spectroscopic properties. The operating wavelengths of many of these compounds, including 3 (excitation 340 nm), fall within the UV region of the spectrum. A major challenge for developing fluorescent sensor devices from these compounds is finding optically compatible sensor materials.40,81-83 Many currently available materials can contribute to interferences and auto fluorescence of a sensor, particularly in the UV region (